Fraunhofer ISI Discussion Papers Innovation Systems and Policy Analysis No. 52

ISSN 1612-1430

Karlsruhe, July 2016

Addressing directionality: Orientation failure and the systems of innovation heuristic.

Towards reflexive governance.

Ralf Lindner, Stephanie Daimer, Bernd Beckert, Nils Heyen,

Jonathan Koehler, Benjamin Teufel, Philine Warnke, Sven Wydra

Karlsruhe, Fraunhofer ISI

Contents I

Contents

1 Introduction .......................................................................................................... 1

2 Conceptual foundations for a revised systems of innovation

heuristic ............................................................................................................... 7

2.1 Systems of innovation approaches and first steps towards re-

conceptualisation ................................................................................. 7

2.2 Grand challenges and mission oriented policies ................................ 11

2.3 STS – micro level thinking ................................................................. 14

2.4 Governance of science, technology and innovation ........................... 14

3 Reflexivity of innovation systems .................................................................... 18

3.1 Reflexivity stages and capabilities ..................................................... 18

3.2 Ten quality criteria for reflexive innovation systems ........................... 22

3.3 A revised heuristic of innovation systems .......................................... 27

4 Discussion and conclusions for STI research and policy .............................. 30

5 Reference List .................................................................................................... 34

II Contents

Figures

Figure 1: Graphical representation of a Reflexive Innovation System (based

on Kuhlmann and Arnold 2001: 2, own compilation) ................................. 28

Figure 2: Improved graphical representation of a Reflexive Innovation

System (own compilation based on Warnke et al. 2016) ........................... 29

Tables

Table 1: Examples of mission oriented innovation policies...................................... 13

Table 2: Quality criteria of reflexive innovation systems .......................................... 26

Introduction 1

1 Introduction

Since about 15 years, science, technology and innovation (STI) policies are increas-

ingly geared towards addressing objectives reaching beyond an immediate economic

focus on growth and competitiveness. This "normative turn" (Daimer et al. 2012) is ex-

pressed in the strategic reorientation of national and supranational STI policies to ad-

dress the so-called ‘grand challenges’ (Kallerud et al. 2013). Well known examples for

this ongoing paradigm shift are the European Union's Europe 2020 strategy, the US

Strategy for American Innovation or Germany's Hightech Strategy. What is more, the

quest to address ‘grand challenges’ such as health, demographic change, wellbeing

and sustainability by the means of research and innovation is complemented by and

propelled forward by the emerging discourse on responsible (research and) innova-

tion.1 In essence, RRI aims at improving the alignment of the impacts of technology

and innovation with societal demands and values as far as possible. The concept is

inherently characterised by a high degree of normativity in order to provide necessary

guidance as to what constitutes desired or ‘responsible’ research and innovation

(Randles et al. 2014; Lindner and Kuhlmann 2016). The prominent position of RRI in

the European Union's research and innovation programme 'Horizon 2020' and the en-

dorsement of the "Rome Declaration on RRI in Europe" by the European Council in

2014 indicate that RRI is increasingly developing relevance for policy, research funding

and scientific communities.

Arguably, the articulation and growing relevance of normative directions of research,

technology development and innovation in addition to and beyond the objectives of

economic growth and international competitiveness signify a paradigm-shift in STI pol-

icy. While this reorientation towards addressing challenges, which can be empirically

observed, might be welcomed from a normative point of view, it poses significant chal-

lenges for the substance, procedural design and coordination of STI. Since the late

1990s, the systems of innovation approach (cf. Edquist 1997; 2005) has established

itself as the most influential paradigm within the international innovation research

communities. The systems of innovation perspective does not only frame the scientific

debates dealing with innovation, it also provides conceptual orientation and strategic

guidance for many governments and international and supranational organisations

such as the OECD and the European Union (Fagerberg and Verspagen 2009; Lindner

2012).

1 For the purpose of this paper, the concepts “Responsible Innovation” and “Responsible Research and Innovation” with its acronym RRI will be used interchangeably. For a discus-sion of the different understandings and applications cf. Tancoigne et al. (2016).

2 Introduction

However, in view of the observed normative turn in STI policy, the innovation system

heuristic is increasingly being criticized for its inability to incorporate the dimension of

normative direction and/or the lack of openness to address objectives beyond systemic

imperfections (Daimer et al. 2012; Weber and Rohracher 2012). The innovation system

approach has mainly focused on the explanation of (cross-national) differences in the

innovation performance of national/technological/sectoral systems by factors outside of

the neoclassical framework (Lundvall 1992; Nelson 1993; Sharif 2006). In the course of

further developments, market failure as the predominant policy rationale of the time

and the chief justification for policy intervention was complemented by the system fail-

ure rationale, which takes into account shortcomings such as infrastructural failures,

capability and learning failures, transition failures, network failures or institutional fail-

ures (Woolhtuis et al. 2005; Chaminade and Edquist 2006; Gassler et al. 2006). Con-

sequently, the innovation system approach is applied with the aim to analyse and

tackle structural issues of an innovation system as a means to foster competitiveness,

economic growth and employment. So far, the innovation system heuristic neither pro-

vides sufficient analytical nor conceptual orientation for questions related to fostering

innovation towards certain ends in addition to and/or beyond the narrow economic

growth paradigm (Quitzow 2011; Bajmócy and Gébert 2014). As a result, established

practices and institutions continue to operate with the chief aim of improving innovation

capacities by "getting the structures right", thereby sidestepping the question "getting

the structures right to achieve what?".

However, despite the fact that the innovation system approach and its numerous vari-

ants have frequently been criticised for, among other weaknesses, their one-

dimensional focus on economic development, the tendency to overemphasize techno-

logical innovation and implicit mechanistic notions (Soete et al. 2010; Dodgson et al.

2011; Mahroum 2012), we claim that the heuristic itself continues to provide useful

analytical lenses and constitutes a valuable conceptual frame of reference for the de-

sign of STI policy. Most notably and well established within the STI community, the

emphasis of the innovation system approach on interactivity and interdependence be-

tween different actors, the understanding of innovation as a collective endeavour which

is influenced by complex framework conditions, the awareness about the roles of dif-

ferent sources of innovation (e.g. non-R&D activities) and the prominent position of

reciprocal learning processes as chief drivers of innovation remain important guide-

posts (Metcalf 2003; Soete et al. 2010).

Given these valuable analytical qualities, the question is raised how the innovation sys-

tem approach should be revised and developed further in order to respond to the chal-

lenges of directionality and normative orientation.

Introduction 3

Reflexive innovation systems

As an answer to this question, we propose to introduce a set of conceptual elements

with the objective of enabling the systems of innovation heuristic to incorporate re-

quirements of directionality and normative orientation. These four elements, constitut-

ing what we term reflexive governance, are conceptualised as quality criteria of innova-

tion systems – understood as ensembles of all relevant institutions and actors involved

in the development, diffusion and use of innovation (Edquist 2005: 182; Kuhlmann et

al. 2010: 3; Warnke et al. 2016: 2). These quality criteria shall help to identify, assess

and ultimately guide innovation processes towards desired directions. The four capaci-

ties for reflexive governance of innovation systems are:

Self-reflection capacities,

Bridging and integration capacities,

Anticipation capacities, and

Experimentation capacities.

Section 3 presents the four capacities in detail, and demonstrates how they relate to

the three stages of ‘directional’ innovation processes – situation analysis, goal formula-

tion and specification, and implementation – which we differentiate for analytical pur-

poses. In fact, ‘directional’ innovation processes are not assumed to be more linear

than ‘undirected’ innovation paths, but the difference is that the concept of reflexive

governance makes the steps of situation analysis and goal formulation/specification

explicit and required parts of the process. The definition of the situation and the formu-

lation of goals (hitherto growth and competitiveness) are thus internalized and no

longer exogenous to the innovation system.

The term reflexivity was chosen to characterise the proposed quality criteria as they

essentially attend to an innovation system's collective ability to reflect about a given

situation, to deliberatively define the goals of innovation and eventually to transpose

these into strategy. Applying the concept of governance in the context of the systems

of innovation heuristic is uncommon. Even so, we suggest to introduce the notion of

governance, understood as the intentional interaction and coordination of state and

non-state actors with the aim to influence innovation systems (Borrás and Edler 2014:

14), in order to conceptually capture actors' purposeful attempts to change decisions

and decision-making processes, framework conditions and STI processes towards cer-

tain ends (or to prevent such change).

To recapitulate, our points of departure for this proposition are (1) the observation of a

conceptual ‘blind spot’ in the established systems of innovation heuristic. While the

systems of innovation approach primarily serves to identify relevant system elements

4 Introduction

and supports the analysis of the interplay of interaction and knowledge exchange, it

fails to provide conceptual underpinnings on the requirements for an innovation system

to identify, assess and ultimately instigate actions with the aim of guiding innovation

towards desired directions. This deficit might also be termed as the 'governance gap' of

the conventional systems of innovation approach. (2) Contemporary literature on the

governance of STI provides a rich reservoir of approaches, instruments and mecha-

nisms on how to steer, influence or 'nudge' actors and institutions with the specific aim

of modulating and orchestrating innovation trajectories according to desired ends.

Hence, we propose to systematically integrate a dedicated governance perspective into

the systems of innovation heuristic in order to address the identified orientation failure.

To a large extent, the rise of the analytical term ‘governance’ since the 1980s is a reac-

tion to the perception of the state's weakening capacity to effectively steer, and to the

growing interdependence between societal subsystems such as different policy areas

and communities or different functional areas. As growing interdependencies in general

mean increasing complexity of policy and politics, this implies that cooperation and

coordination between different actors become progressively more important (Jessop

2003: 103). Given the highly complex nature of the grand societal challenges, which

require far-reaching socio-economic transformation if they are to be adequately ad-

dressed (Kuhlmann and Rip 2014; OECD 2015), governance approaches suggest

themselves as a potentially useful contribution when dealing with complex actor con-

stellations and their interplay.

The rationale for introducing the element of reflexivity in the re-conceptualisation of the

systems of innovation approach, operationalised as a set of four capacities, is twofold.

First and most obvious, a conceptual vehicle was sought to integrate substantial issues

(normative direction, problem-/solution-orientation, policy goals, missions etc.) in sys-

tems of innovation thinking. To this end, the governance perspective with its analytical

focus on actors' intentionality to influence innovation and its fittingness to cope with

interdependencies was proposed. Second, such a rather technocratic approach is

prone to advance or at least assert top-down prescriptions which normative directions

an innovation system should take. ‘Grand challenges’, missions or similar broad and

far-reaching substantive goals and related implementation measures are likely to face

low acceptance and will ultimately suffer under weak legitimacy when they are not well-

grounded in broad-based processes of collective sense-making, deliberation and nego-

tiation. Thus, we propose to endogenise the processes of defining and concretising

normative directions and their implementation within an innovation system. While the

broad objectives and directions to take remain to be set through formally legitimised

representative-democratic decision-making routines, the STI actors need to be ade-

quately involved in the concretisation of directions and the definition of socio-technical

Introduction 5

pathways. In order to achieve this, innovation systems require reflexive capacities and

appropriate governance mechanisms.

In addition to the objective of addressing orientation failure, the emphasis on reflexive

capacities in the systems of innovation framework also resonates with and responds to

a number of interrelated phenomena and developments observed by contemporary

innovation research:

As Warnke et al. (2016) convincingly show, the actor landscape in innovation sys-

tems has become significantly more diverse compared to the established under-

standing reflected in systems of innovation literature, calling for a broadened view of

the relevant actors as well as a revision of their relationships and the functions as-

signed to them.

Closely related to the appearance of new actors in innovation systems, new types of

innovation, such as user-driven innovation (von Hippel 2005), social innovation

(Moulaert et al. 2013) or forms of collaborative innovation (Warnke et al. 2016: 10 ff.),

are increasingly recognised as relevant, thus changing the dynamics of innovation

systems accordingly.

Science-society relations are shifting, indicated by changing expectations of society

towards science and research concerning contributions to societal challenges and

value-orientation. Moreover, knowledge availability and science-literacy have in-

creased, empowering parts of the citizenry to critically assess and proficiently ques-

tion science and research.

STI and related policy are facing increased pressures to legitimise their actions and

outcomes, calling for more transparency and higher levels of readiness of scientists

and researchers to engage with and respond to stakeholders and lay people.

The increasing influence of societal missions and challenge-oriented research has

sparked debates about the status of academic freedom and scientific autonomy. Ar-

guably, the collective processes of goal formulation in reflexive innovation systems

have the potential to regain some influence for research and innovation actors con-

cerning which directions to take, when they are acting with an outward-looking per-

spective. And in a reflexive innovation system, the balance between investigator-

governed research and mission-governed research (Arnold and Giarracca 2012)

should be subject to deliberate social choice.

Participatory approaches and particularly upstream engagement as important ap-

proaches to social control of technology have received much attention in STI policy

discourse (Wilsdon and Willis 2004; Fisher et al. 2006). However, they have also re-

ceived criticism for implicit notions of linearity, technology determinism and their vul-

nerability to instrumental use (Stirling 2008: 264). The proposed reflexive capacities,

which reach beyond participatory mechanisms, address all phases of the innovation

process and deliberately incorporate processes of deliberation, experimentation and

also policy or "strategic intelligence" (Kuhlmann et al. 1999). They can also be un-

6 Introduction

derstood as an attempt to breakup these confines and avoid prematurely closing

down the possible range of technological choice (Stirling 2008). This connects to the

STS literature and its Social Construction of Technology framework. If technology is

socially ‘constructed’, then the goals and norms of society will necessarily play a

role.

Based on the considerations outlined in this introduction, we propose that reflexivity or

reflective capacities should be major criteria of the quality of innovation systems. The

ability of innovation systems to collectively reflect about orientation and objectives, and

to bring together actors to deliberate goals, to prioritize and to specify them should en-

ter the systems of innovation concept as quality criteria in addition to the system func-

tionalities currently primarily endorsed by the concept. Taken together, the reflexive

capacities constitute an additional 'reflexivity layer' in the innovation system, and func-

tions as a framework condition similar to the socio-cultural context of a system (see

also section 3.3).

In the following section 2, the development of the main strands of the systems of inno-

vation literature will be discussed and complemented with the key insights of contem-

porary debates related to the ‘grand challenges’ and the governance of STI. In section

3, our proposition to revise the systems of innovation approach by incorporating reflex-

ivity and associated quality criteria is presented in detail. The final section 4 discusses

the conceptual contribution and draws conclusions for STI analysis and policy.

Conceptual foundations for a revised systems of innovation heuristic 7

2 Conceptual foundations for a revised systems of innovation heuristic

2.1 Systems of innovation approaches and first steps to-wards re-conceptualisation

As briefly outlined above, the systems of innovation heuristic (Lundvall 1992; Nelson

1993; Edquist 1997) continues to be the most influential paradigm within the interna-

tional innovation policy community. In systems of innovation thinking, innovation is

conceptualised as a nonlinear, evolutionary and interactive process characterised by

reciprocity and iterative feedback mechanisms, in which actors (e.g. firms), organisa-

tions (e.g. universities, customers, government) and institutions (e.g. regulations, cul-

ture) interact in many ways. A distinction can be made between national (Freeman

1995; Nelson 1993), regional (Braczyk et al. 1998; Cooke et al. 1997), sectoral (Bre-

schi and Malerba 1997; Malerba 2002) and technological innovation systems (Carlsson

1995; Carlsson and Jacobsson 1997). Particularly the latter has increasingly become a

popular concept in analyzing innovation processes concerned with emerging technolo-

gies. Technological innovation systems (TIS) are defined as "socio-technical systems

focused on the development, diffusion and use of a particular technology (in terms of

knowledge, product or both)" (Bergek et al. 2008: 408). The TIS framework consists of

different structural elements, such as actors or networks of actors, institutions and the

interactions between them (Markard and Truffer 2008; Wieczorek and Hekkert 2012).

Next to structural components, the focus on systemic processes (functions) within TIS

is a more recent addition to the concept (e.g., Hekkert et al. 2007; Suurs et al. 2009;

van Alphen et al. 2010). Hence, within the TIS framework, the structures and functions

are usually analyzed on a system level.2

Innovation system analysis mainly focuses on the explanation of the economic per-

formance of national/regional/technological/sectoral systems by factors not considered

by mainstream neoclassical thinking in economics (Sharif 2006). The predominant pol-

icy rationale of market failure has been complemented by the system failure concept,

which takes into account weaknesses such as infrastructural failures, capability and

learning failures, transition failures, network failures (weak and strong) or institutional

failures (Woolthuis et al. 2005; Chaminade and Edquist, 2006). Thus, the systems of

innovation approach is usually applied as an alternative to the neoclassical approach

2 The literature usually operates with seven key systems functions which technological inno-vation systems need to fulfil: 1. entrepreneurial experimentation, 2. knowledge development, 3. knowledge exchange, 4. guidance of the search, 5. formation of markets, 6. mobilization of resources, and 7. counteracting resistance to change (Hekkert et al. 2007).

8 Conceptual foundations for a revised systems of innovation heuristic

with the aim to tackle generic questions of innovation and its role for economic growth

and employment. To a large extent, this narrow focus on economic growth and em-

ployment has been reinforced by corresponding demands of policy makers. Edquist

points out for example: "I assume that the objectives – whichever they are – are al-

ready determined in a political process. It should also be mentioned that they do not

necessarily have to be of an economic kind. They can also be of a social, environ-

mental, ethical or military kind. They must be specific and unambiguously formulated in

relation to the current situation in the country or in other countries. With regard to inno-

vation policy the most common objectives are formulated in terms of economic growth,

productivity growth or employment." (2002: 220). Also in Porter's work (1990), the role

of policy in directing innovation is mentioned, however this does not refer to strategic

priority-setting but rather to standards and regulations. Apart from that, Porter and Nel-

son (1993) agree in their view that the direction of innovation is a function performed

mainly by customers.

However, it has frequently been criticized that innovation system analysis places too

much emphasis on the strengthening of generic innovation capabilities and less on

certain outcomes (Wydra 2015). As Daimer et al. argue: "Despite all the refined under-

standing of innovation systems, the instruments derived from the innovation system

approach are mainly directed at enhancing the innovation ecosystem in order to

strengthen innovation capability. So far, there is no attempt to build on the innovation

system heuristic in order to modulate innovation journeys towards certain desirable

objectives. So whereas system failure appears to be addressed, 'orientation failure' has

largely not been tackled." (2012: 222). Similarly, Quitzow points out: "To date, the sys-

tems of innovation approach has largely been employed to tackle generic questions of

innovation and its role in economic development. From a policy perspective, research

has not focused on understanding how policy might steer innovation in a particular di-

rection (i.e. towards improving the environmental performance of the economy) but on

how to strengthen innovation within the economy more generally." (2011: 13).

Criticism has also been raised against the TIS approach. This concept presumes that

the diffusion of a pre-defined specific technological solution is desirable from a socio-

economic point of view. This perspective has its limits particularly in a long-term per-

spective, given the complexity and uncertainty of innovation paths. In addition to the

problematic ex-ante determination of a technological choice, Truffer points out that "TIS

scholars are blamed to reduce transitions to a simple problem of diffusing new and better

technologies, whereas the reorientation of user practices, power relationships, regulatory

structures, mind sets and public discourses remains unaddressed." (2015: 65).

Conceptual foundations for a revised systems of innovation heuristic 9

Mazzucato (2015a, b) arrives at similar conclusions regarding directionality failures in

her current work on the role of the state in innovation processes. She points out that

markets are "blind" (Nelson and Winter 1982; Dosi 1982) and that the direction of

change provided by markets often represents suboptimal outcomes from a societal

point of view. Hence, the market failure approach but also the systems of innovation

heuristic "cannot explain and justify the kinds of transformative mission-oriented in-

vestments that in the past picked directions, coordinated public and private initiatives,

built new networks, and drove the entire techno-economic process, which resulted in

the creation of new markets, not just in the fixing of existing ones" (2015a: 5). With re-

gard to addressing societal challenges, governments have to lead the process and

provide the direction toward new "techno-economic paradigms" (Perez 2002), which do

not come about spontaneously out of market forces. Besides referring to the theoretical

conception of techno-economic paradigms, Mazzucato underpins her argumentation

with the empirical observation that for many technological transformations, govern-

ments made direct "mission-oriented" investments in the technologies that enabled

these revolutions to emerge (Mazzucato 2015a: 6).

Against the background of these discussions and observations, innovation scholars

started to recognize the need to further develop the systems of innovation approach

and related innovation policy concepts in order to better address socio-economic goals.

Bajmócy and Gébert (2014) combine the systems of innovation approach with the ca-

pability approach that has been developed by Amartya Sen. They take the capability

approach as a starting point for an alternative to the growth-centred paradigm and ex-

amine whether such a shift to well-being as the main policy objective changes the set

of information needed for the design, implementation and evaluation of innovation poli-

cies. They conclude that the capability approach would broaden the boundaries of in-

novation system analysis and requires a change in the informational basis. However,

the authors do not focus on the operationalisation of such an extended innovation

model.

Weber and Rohracher (2012) integrate insights from the systems of innovation ap-

proach and a Multi-Level Perspective (MLP) (Geels 2002) in a comprehensive 'failures'

framework that includes the focus on societal goals. They claim that the combination of

MLP's goal-oriented system transformation approach, complemented by the related

approaches of Strategic Niche Management (Kemp et al. 1998) and Transition Man-

agement (Rotmans and Loorbach 2006), with structure-oriented systems of innovation

approaches improve the conceptual foundation and actual implementation of transfor-

mation oriented innovation policies. The authors propose the recognition of additional

types of failures next to market and system failures in an innovation (production and

consumption) system, the so-called transformational system failures in order to take

10 Conceptual foundations for a revised systems of innovation heuristic

into account the requirements of goal-oriented transformative change. Among others,

they identify two types of failures that seem to be linked to socio-economic outcome

orientation: directionality failure and reflexivity failure. Regarding directionality failure

they point "… to the necessity not just to generate innovations as effectively and effi-

ciently as possible, but also to contribute to a particular direction of transformative

change. This direction is defined, for instance, by the identification of major societal

problems or challenges, for which solutions need to be developed with the help of re-

search and innovation" (Weber and Rohracher 2012: 142).3 However, different failure

mechanism may occur, e.g., the lack of a shared vision regarding the goal and direc-

tion of the transformation process or the inability of collective coordination of distributed

agents involved in shaping systemic change. Concerning reflexivity failure, they claim

that "…a continuous monitoring with respect to progress towards the transformation

goals and the development of adaptation strategies" (Weber and Rohracher 2012: 142)

is required to address the uncertainty surrounding innovation and change in the trans-

formative process. Potential failure mechanisms may relate e.g. to a lack of distributed

reflexive arrangements to connect different discursive spheres, and insufficient spaces

for experimentation and learning (Weber and Rohracher 2012).

Apart from these contributions concerning the rationale of innovation policy, a few au-

thors analyze potential implications for governance and policy measures to better

achieve such socio-economic goals4.

Daimer et al. (2012) propose that systems of innovation approaches should incorporate

an orientation function as an integral element in order to improve innovation capabilities

to address ‘grand challenges’. They study the refining of policy instruments towards

addressing societal outcomes for two systemic policy instruments, evaluation and fore-

sight. For example, participatory evaluation approaches could be complemented with

the analysis of new impact types (e.g. sustainability) or behavioural additionality. Simi-

larly, foresight processes that explore innovation journeys as an element of challenge

oriented innovation policy strategies are identified as suitable orientation instruments.

Mahroum (2012) claims that most existing innovation policy tools only address con-

straints and barriers of innovation linked to broader socio-economic problems, but the

policy instruments themselves are not linked to the desired outcome. In order to estab-

lish those linkages, he introduces an analytic-diagnostic framework that aims to help

develop innovation policies designed to achieve certain socio-economic outcomes. The

3 This line of reasoning may be combined with rationales of other aspects of mission-oriented policy, such as policy coordination etc.

4 The review is based on earlier work of Wydra (2015).

Conceptual foundations for a revised systems of innovation heuristic 11

main idea is to start from the desired outcome and deduct (among others) policy ra-

tionales and policy instruments. Overall he identifies four categories of outcomes which

would be addressed by different interventions (e.g. mission-oriented policies): develop-

ing capabilities to meet critical needs in defence, environment and health; broad and

diversified capabilities to create higher living standards; generation of new supply

chains that would create new value; wide market uptake for specific solutions such as

alternative energy products.5

To conclude, there have been different contributions to the justification of mission-

oriented policies as well as some approaches to develop innovation policies to achieve

certain socio-economic outcomes. However, they do not systemically elaborate how

the innovation system concept could be modified to be able to analyse orientation ca-

pabilities and thus contribute to the avoidance of orientation or outcome failures.

2.2 Grand challenges and mission oriented policies

Regarding policy practice, the notion of ‘mission orientation’ in STI policy was subject to

significant transformations since the mid of the 20th century. Daimer et al. (2012: 218 ff.)

outline several paradigm shifts of innovation policy. In the 1960s, the primary goal of

innovation policy was to fix ‘market failures’ by funding basic research. This was fol-

lowed by several different phases of mission oriented innovation policies that were de-

signed to reach pre-defined goals: After the first phase of ‘classic’ mission orientation

that was mainly characterized by funding schemes such as the US Apollo 1 space pro-

gram, the second phase that started in the 1980s combined several policy instruments.

In this second phase, non-linear, recursive interactions of heterogeneous actors were

addressed, however, "policies solely targeted selected sectors and technologies"

(Daimer et al. 2012: 219). In the 1990s, innovation policy entered a third phase that

was driven by the idea of optimising the ‘innovation ecosystem’ to improve innovation

capabilities with the aim to strengthen competitiveness and economic growth. Policy

instruments were targeted at enhancing systems' learning capability, improving the

management of interfaces and capacity building among different actors in the innova-

tion system. The application of innovation system thinking and analysis did not neces-

sarily come along with new policy paradigms as they rather complemented already

established approaches.

5 However, he only links broad strategies of innovation policy making to certain goals and does not specify to which extent certain instruments themselves change. Moreover, the possibility that one policy intervention may intend to address several potential goals at the same time is not considered.

12 Conceptual foundations for a revised systems of innovation heuristic

Today, the rationales of competitiveness and economic growth have been comple-

mented by additional goals of STI policy that can be summarized as addressing ‘grand

challenges’, such as adaption to climate change, health or integrative societies. This

transformation of goals beyond a quantitative economic output is what Daimer et al.

(2012: 218 f.) call a "normative turn" in innovation policy and related innovation activi-

ties. The corresponding idea of a new mission orientation typically combines several

characteristics (cf. Dachs et al. 2015: 5 ff.): In order to achieve societal goals, transdis-

ciplinarity, international collaboration, and emphasises a broad dissemination of inno-

vation need to be fostered. Its instruments are open to different competing technolo-

gies, and focuses on system innovations that integrate both social and technological

innovations. It addresses a wider circle of actors within the innovation system, and is

systematically coordinated between different ministries, departments and other public

agencies at different levels.

An empirical comparison of today's national innovation policies in different countries

(OECD 2014: 110 ff.) reveals significant differences in the realization of this new mis-

sion orientation: While many national innovation strategies lack an explicit reference to

‘grand challenges’, most of the other strategies are designed "[…] both to strengthen

growth and to address a range of global and social challenges, including climate

change and health." (OECD 2014: 94). Table 1 shows all national and supranational

innovation policies/strategies that are explicitly related to ‘grand challenges’. The ‘grand

challenges’ addressed across these strategies can be attributed either to a societal

domain or to sustainability/‘green growth’. Most of these strategies address several

‘grand challenges’ ranging over different domains. Remarkably, they often represent

issues that are related to specific characteristics of the respective country. Thus,

speaking of ‘global grand challenges’ seems exaggerated at this point.

The French National Research Strategy represents an innovation policy that is already

quite similar to the idea of a new mission orientation described above: French industrial

policies have always been mission-oriented, they were driven for example by military

missions, but also by grand societal challenges, such as public health. The current

strategic turn was initiated in 2013 by president Hollande and is motivated by an endur-

ing economic crisis and the need for ‘structural’ economic and social reform. Strategy-

making involved a forward-looking approach in order to address major transformations

under way (ageing, digital revolution, transition to low-carbon economy, metropolisation

and more), which call for programming of government action (Pisani-Ferry 2015). Next

to the anticipatory approach, the strategy has strong learning and participatory ele-

ments.

Conceptual foundations for a revised systems of innovation heuristic 13

Table 1: Examples of mission oriented innovation policies; source: OECD

(2014)

Country/strategy Goals/mission orientation

Argentina/Bases for an STI Strategic Plan

Increase consistency and social equality, Promote sustainable development

Brasil/National Strategy for Science, Technology and Innovation (ENCTI)

Promote a green economy; contribute to eradicating poverty and decreasing social and regional inequalities

China/Medium and Long-term National Plan for Science and Technology Development

Build a conservation-minded and environmentally friendly society

Columbia/Sectoral Strate-gic Plan for Science, Technology and Innovation

Promote knowledge and innovation for production and social transformation

Japan/4th S&T Basic Plan Comprehensive promotion of S&T and innovation and an is-sue-driven approach through: integrated development of STI policies to address societal challenges; realisation of a policy to be created and promoted with civil society. Priority areas: environment; energy; health and medi-cal/nursing care; social challenges.

France/National Research Strategy (SNR)

Ten societal challenges are identified, and for each challenge a research strategy is defined; a strategy for large equipment, a limited number of major scientific and technological priorities and some steering rules. The ten challenges: sustainable resource management and adaptation to climate change; safe, effective and clean en-ergy; stimulate industrial revival; health and wellness; food security and demographic challenge; sustainable mobility and urban systems; information society and communication; inno-vative, integrative and adaptive societies; spatial ambition for Europe; freedom and security for Europe, its citizens and its residents.

South Africa/National De-velopment Plan (NDP): A Vision for 2030

Give South Africa a diversified economic base by extracting more local value from mineral resources, ensuring access to good quality water and alternative sources of energy, identify-ing new and innovative ways to address poverty, inequality and the burden of disease. Priorities areas: water, power, marine, space and software engineering.

EU28/EU Framework Pro-gramme for Research and Innovation – Horizon 2020

Meeting societal challenges: address concerns of citizens in Europe and elsewhere (health and well-being, food security, sustainable agriculture, bioeconomy, secure and clean en-ergy, smart and integrated transport, environment, resource efficiency, inclusive, innovative and secure societies).

14 Conceptual foundations for a revised systems of innovation heuristic

2.3 STS – micro level thinking

Science and technology studies (STS) are the study of the processes and outcomes of

science and technology. STS is interdisciplinary, including sociology, history, philoso-

phy of science and technology, and anthropology (Sismondo 2010). For the purposes

of governance of innovation, we summarise the technology side of STS. It views tech-

nology as well as science as social activities. It is non-positivist, in particular, it argues

that there is no technological ‘method’ that can simply be used to apply scientific

knowledge to produce technological artefacts.

"...the interpretations of knowledge and artefacts are complex and various: claims,

theories, facts and objects have very different meanings to different audiences."

(Sismondo 2010: 11)

Instead, both science and technology are ‘constructed‘ by people in society. Both sci-

ence and technology are therefore products not only of objective laws, but of their so-

cial context. This has led to the idea of Social Construction of Technology (SCOT).

There is little direct consideration of the governance of innovation in the basic introduc-

tion to STS in Sismondo (2010). However, science policy is also a result of the insights

generated by STS.

This constructivist approach to technology implies that technology is influenced by pol-

icy as one part of society, together with other social actors. This then is similar to the

innovation systems analysis, but with an emphasis on the sociological aspects of tech-

nology – how it plays a role in peoples’ lives and how actors understand their use of

artefacts to fulfil their needs.

The implication for reflexive governance is quite strong: innovation is a social phe-

nomenon, determined not just by the scientific and empirical knowledge in society, but

also by the views and needs of social actors. Governance processes can therefore play

a role in determining and realising the direction of innovation, as can the other actors

involved in technological development.

2.4 Governance of science, technology and innovation

The debates on governance of STI are largely detached from conceptual developments

of the systems of innovation heuristic. However, against the background of the chal-

lenges associated with the growing importance of directionality and normativity in STI,

governance and its actor perspective are becoming more important as a means of pro-

viding orientation, guidance and eventually coordinating/orchestrating the complex

processes of societal transformation.

Conceptual foundations for a revised systems of innovation heuristic 15

While a broadly shared and uncontroversial definition of governance does not exist,

most would agree that the concept refers to the increased role of non-government ac-

tors in policy-making (Bache 2003). Most importantly, forms of hierarchical, top-down

command-and-control approaches have been increasingly complemented and often

supplanted by non-hierarchical forms of decision-making such as negotiation, consulta-

tion or dialogue (Jessop 2003: 104). In this sense, governance can be located some-

where between the market's 'invisible hands' and the 'iron fist' of centralised, hierarchi-

cal government (Jessop 2003: 101). In an overview of the debate, Grande (2012: 566 f.)

identifies five core elements which seem to constitute the core of contemporary under-

standings of governance: (1) importance of non-hierarchical forms of decision-making;

(2) a growing role of non-state actors; (3) growing interdependencies between policy

areas and societal subsystems; (4) increasing complexity; and (5) increasing impor-

tance of coordination and cooperation.

The rise of the term governance during the last three or four decades is related to a

number of interconnected changes in the way contemporary, highly industrialised so-

cieties are organised and governed. In essence, governance can be understood as a

response to intensified social complexity. According to Scharpf, "the advantages of

hierarchical coordination are lost in a world that is characterized by increasingly dense,

extended, and rapidly changing patterns of reciprocal interdependence, and by increas-

ingly frequent, but ephemeral, interactions across all types of pre-established bounda-

ries, intra- and interorganizational, intra- and intersectoral, intra- and international"

(1994: 37).

However, the increased importance of non-state actors – a development often referred

to as "from government to governance" – does not simply result in a diminishing role

for government per se. Instead, the changes in the modes of coordination we can ob-

serve are not a uniform shift away from government, but represent multiple and inter-

twined changes between state intervention and societal autonomy in which different

forms of rule setting, steering and governing co-exist and interact (Jordan et al. 2005:

484; Mayntz 2009: 105; Lange et al. 2013).

Many definitions of governance are based on the key elements outlined above and

thus emphasis – in a variety of combinations and by stressing different aspects – the

interactive processes of purposeful coordination and/or management between different

(types of collective) actors in coupled and overlapping arenas (cf. Kooiman 1993;

Kuhlmann 2001; Jessop 2003; Benz 2006; Voß et al. 2006; Mayntz 2009).

16 Conceptual foundations for a revised systems of innovation heuristic

With the aim of making the general governance concept fruitful for our conceptual ap-

proach to reflexive innovation systems, we draw on Borrás and Edler (2014) who de-

fine governance in relation to STI as the

"[...] way in which societal and state actors intentionally interact in order to transform

ST&I systems, by regulating issues of societal concern, defining processes and di-

rection of how technological artefacts and innovations are produced, and shaping

how these are introduced, absorbed, diffused and used within society and econ-

omy." (2014: 14).

Two aspects make this definition particularly useful compared to other suggestions.

First, Borrás and Edler (2014) provide a definition which specifically refers to research

and innovation systems, thereby highlighting the complex processes of co-construction

and diffusion. Second and even more important in our context, the definition explicitly

addresses actors' purposeful attempts to influence decisions and decision-making

processes, framework conditions and STI processes towards certain ends (or to pre-

vent such change). This understanding also underscores the fact that governance

more often than not encounters contestation between different problem definitions,

goals, and interests.

The topic of reflexivity or reflexive governance is not new in STI-related research. On

the contrary, an impressive number of approaches and concepts dealing with questions

of societal embedding of technology and innovation, steering socio-technical change,

and the integrated assessment of the potential social, economic and ecological impacts

of STI have been developed and implemented in the past. To some extent they consti-

tute what might be termed 'de facto reflexive governance'.6

One of these concepts is 'strategic intelligence', understood as a set of sources of in-

formation and explorative as well as analytical (theoretical, heuristic, methodological)

tools employed to produce 'multi-perspective' insight in the actual or potential costs and

effects of public or private policy and management (Kuhlmann et al. 1999). Typically,

Technology Assessment (TA), foresight processes, evaluations and the like constitute

strategic intelligence and share the objective of contributing to the development of STI

strategies by analysing and assessing their broader implications in advance (Ely et al.

2014; Lindner et al. 2016). While strategic intelligence is primarily – but not exclusively

– based on expert knowledge and intends to enlighten and rationalise debate and deci-

6 Using the term ‘de-facto reflexive governance’ in this context was originally inspired by Rip’s (2010) discussion of ‘de facto governance’. Its application was further elaborated by Randles et al. (2016) who refer to governance related to responsible research and innova-tion.

Conceptual foundations for a revised systems of innovation heuristic 17

sion-making, other governance approaches place less emphasis on appraisal and

rather focus on the active integration of diverse perspectives and stakeholders in the

actual processes of STI. Well-known examples include Constructive TA (Rip et al.

1995), real-time TA (Guston and Sarewitz 2002), mid-stream modulation (Fisher et al.

2006), value-sensitive design (van der Hoven and Manders-Huits 2009) or anticipatory

governance (Guston 2014). Most of the rather recent conceptual contributions to the

RRI debate follow this line of reasoning by explicitly emphasising reflexivity in the gov-

ernance of research and innovation (cf. Stilgoe et al. 2013; Kuhlmann et al. 2016).

These de facto reflexive governance approaches can be seen as a response to the

increasing speed, dynamic and uncertainty of contemporary technology development.

The approaches' common objective to modulate STI trajectories in a forward-looking,

anticipatory and broad-based manner integrating multiple perspectives is driven by the

hope to come to terms with the Collingridge-Dilemma (Collingridge 1980). The funda-

mental uncertainties we face when dealing with STI is further amplified by the complex-

ity and ambivalence associated with the ‘new missions’ or societal challenges ad-

dressed by contemporary STI policy as they require broad systemic transformations

encompassing interdependent social, technological, economic and ecological ele-

ments.

Against this background, approaches are needed which take into account complexity,

uncertainty and ambiguity. Reflexive governance currently seems to be the most ap-

propriate governance mode in the field of STI due to its openness towards alternative

solutions and pathways, experimentation and learning. Instead of trying to follow the

illusion of complete knowledge and control, reflexive governance fosters continuous

reflection and learning while tentatively modulating developments (Voß et al. 2006: 6 f.).

18 Reflexivity of innovation systems

3 Reflexivity of innovation systems

In our understanding, reflexivity of innovation systems means a set of system qualities

and processes underpinning the ability to address directionality in innovation. Reflexiv-

ity denotes a specific quality of an innovation system, similar as framework conditions

such as culture, infrastructure, standards, norms or policies defining the 'rules of the

game'. Reflexivity is a 'layer' of an innovation system which is embodied in certain ca-

pabilities owned by individual actors in the system, by the organizations (collective ac-

tors) in the system, and hence by the system as a whole (see section 3.3). These ca-

pabilities are reflected by the attitudes and expressed opinions of individual and collec-

tive actors and by the way they organize processes and interactions in the system. To

illustrate the abstract concept, we provide dimensions of reflexivity, indicators and ex-

amples below.

We propose reflexivity as a response to increasing complexity in innovation because

an innovation system with reflexive capabilities is able to better endogenize the direc-

tionality question. Reflexive innovation systems, as we conceptualize them, are charac-

terized by the internalization of problem-oriented and strategic thinking into innovation

processes, by the ability of actors to debate challenges, to agree which challenges

need prioritization and how to concretise their implementation. Thus, reflexivity is an

answer to the currently observed normative debate in innovation policy, in which policy

objectives beyond immediate economic growth and competitiveness – such as ad-

dressing the grand societal challenges of our time – become relevant. This sheds a

different light on how innovation systems can perform the ‘guidance of the search’ –

one of the seven functions put forward in the context of technological innovation sys-

tems (TIS) (cf. section 2.1). TIS-approaches often implicitly assume the ‘guidance of

the search’ to be exogenously set by the decision to follow a certain, pre-defined tech-

nological path. In contrast, a reflexive ‘guidance of the search’ includes collective strat-

egy processes of the system actors to agree on common goals and to pathways on

how to achieve them (and eventually this might mean to pursue multiple (technological)

paths in parallel in the beginning when uncertainty about their contributions to reaching

the common goals is still high).

3.1 Reflexivity stages and capabilities

How do we make the reflexivity of innovation systems more tangible? How can we de-

tect reflexivity in an innovation system? Or what exactly would be needed to reach

higher levels of reflexivity in an innovation system? We have said above that reflexivity

means a set of system qualities which are reflected in capabilities, interactions and

processes. As such, reflexivity is a quality criterion of the innovation system.

Reflexivity of innovation systems 19

In this section, we define a set of quality criteria of reflexive innovation systems. We

have deducted this set of overall ten criteria from literature on innovation systems, in

particular on systemic instruments, on the governance of STI, and based on insights

from sustainability transition literature. This set of quality criteria shall help to implement

our concept in the empirical study of innovation systems as the individual quality crite-

ria can be directly linked to functional units, organizations, and/or policy instruments in

the system.

We have argued that the innovation system heuristic needs a governance perspective,

and that the governance of an innovation system should serve to direct the system. In

order to allow for system-internal strategy formulation and implementation, substantial

orientation of the system needs to be endogenized. In their book "Reflexive Govern-

ance", Voß et al. (2006) differentiate between two meanings of reflexivity. Our under-

standing of reflexivity is very much the same in what they term "second order reflexiv-

ity" – a conscious and strategic way of decision-making and action (2006: 6).

Furthermore, the authors of "Reflexive Governance" define "strategy elements" of re-

flexive governance and locate these along the stages of "system analysis", "goal for-

mulation" and "strategy development and implementation" (Voß et al. 2006: 18). These

stages are consistent with what we argue are the most pressing needs of innovation

systems: A conscious view on system outputs and the kind of directions they serve, as

well as the ability to communicate about goals, priorities and strategies within the sys-

tem. Consequently, the stages "system analysis", "goal formulation" and "strategy de-

velopment" help to ‘politicize’ the innovation system. Instead of keeping overarching

objectives (be it economic growth by way of strong innovation output or the reference

to grand societal challenges as defined by policy) exogenous to the innovation system,

these stages help to internalize related debate. This is important because the overall

strategic orientation of the system is an essential framework condition of all system

activities. In order to avoid 'window dressing' of system outputs, understood as a sim-

ple 're-labeling' of outputs in the language of an overarching strategy, the actors in the

system should be committed to the overall strategy. This is less likely if strategy devel-

opment is external to the system. Thus, the actors in the innovation system should be

able to participate in the making of the strategy. However, the governance of innovation

systems should not be misunderstood as a "self-steering of society" (Voß et al. 2006:

8), but as something in which democratic policy-making of representative systems

plays a central role. These reflexive elements, which allow for participation and interac-

tion of the actors performing research and innovation as well as user groups, are com-

plementary to policy-making processes in representative political systems – and not a

replacement. Or phrased differently, our approach calls for an enhanced role for gov-

20 Reflexivity of innovation systems

ernment to engage in actively shaping and supporting the political and cultural project

of socio-technical transformation (OECD 2015).

The three stages of reflexivity, where these actors need to act and interact, are (own

compilation based on Voß et al. 2006: 18):

System analysis – defined as the identification of needs, demands, values, actors,

structures and trends.

Goal formulation – defined as the definition of interests, positions, goals and visions.

We adjust the definition of this stage and call it goal formulation and specification.

This underlines the need that goals should not only be fixed in general terms (e.g.

sustainability) but most often require further specification.

Strategy development and implementation – defined as the making of a joint strat-

egy and instruments for its implementation and evaluation.

Given the richness in the literature on reflexive governance elements, we identify and

define four types of reflexivity capacities of innovation systems. In the list below, we

show how we combine the various sources in the literature into our scheme of capaci-

ties. The most important sources are Voß et al. (2006) who talk about strategy ele-

ments of reflexive governance, and Weber and Rohracher (2012) who list capacities

which are missing in the innovation system. They group these (missing) capacities into

four types of failures: reflexivity failures, directionality failures, demand articulation fail-

ures and policy coordination failures. As our concept of reflexivity is defined broader,

we view all of these capabilities relevant and include them in the list below.

So far we have spoken of system qualities, but most of these criteria cannot be meas-

ured directly at a system level. While some criteria require central (governmental)

processes or instruments, many others should be present as cultural, structural or pro-

cedural elements in the organizations performing research and innovation as well as in

stakeholder groups. Here it becomes evident why the reflexive innovation system con-

cept is in need of a governance perspective as it entails a dedicated conceptual ap-

proach to the actors in the system. This is needed to understand the actors – be it indi-

vidual or collective ones – not as functional units, but as acting ones, based on their

values, goals and strategies.

Four capacities of reflexivity:

Self-reflection capacities: These can mainly be found at the level of (individual or

collective) actors. They include the critical reflection about values and orientation (cf.

"value debate" in Daimer et al. 2012: 181), "demand-articulating competencies"

(Weber and Rohracher 2012: 1045), as well as the ability to adapt own positions

and goals.

Reflexivity of innovation systems 21

Bridging and integration capacities: Here, we distinguish between capacities at actor

level and capacities at system level. At actor level, bridging and integration capaci-

ties are the (individual) ability (or organizational culture) to think and work transdis-

ciplinary, the openness for different and new knowledge sources, and conflict rec-

ognition and moderation. At system level, bridging and integration capacities can be

found in "arrangements to connect different discursive spheres" (Weber and

Rohracher 2012: 1045), in "joint learning processes between technological AND so-

cietal innovators" or "bi-directional" stakeholder participation (Daimer et al. 2012:

177, 180), in processes or structures that allow for "participatory goal formulation"

(Voß et al. 2006). A "shared vision regarding the goal and direction of the transfor-

mation process" (Weber and Rohracher 2012: 1045) is desirable, but bridging and

integration capacities can also be found in the avoidance of "premature closure on

easy fixes and actively bring these conflicts into the open" (Daimer et al. 2012: 181).

Further system capacities include the "interactive strategy development" (Voß et al.

2006) or the "collective coordination of distributed agents involved in shaping sys-

temic change", "concerted policy initiatives" (Kuhlmann and Rip 2014) as well as

multi-level, horizontal, vertical and temporal "policy coordination" (Weber and

Rohracher 2012: 1045).

Anticipation capacities: These capacities are those of (individual or collective) actors

as well as of the system itself. They include futures literacy (Miller 2007) which can

be the "anticipation of long-term effects" (Voß et al. 2006), the anticipation of trans-

formation (Daimer et al 2012: 181; Kuhlmann and Rip 2014) or the ability to deal

with uncertainty, e.g. by "adaptive policy portfolios to keep options open" (Weber

and Rohracher 2012: 1045). They also include demand literacy, which means "an-

ticipating and learning about user needs to enable the uptake of innovations by us-

ers" (Weber and Rohracher 2012: 1045).

Experimentation capacities: Here, we are thinking of collective experiments (Daimer

et al. 2012: 181; Joly et al. 2010) and thus phenomena which can be found primarily

at system level. We define experimentation as allowing for parallel approaches and

learning through failure on all levels and in different contexts. The need for experi-

mentation has also been highlighted by others, e.g. Voß et al. (2006: 433 f.) call for

"Experimentation and adaptivity of strategies", Weber and Rohracher (2012) for

"spaces for experimentation and learning" and Kuhlmann and Rip (2014) for "tenta-

tive policy mixes". In 2008, the Lisbon Expert Group arrived at the conclusion that

formats such as OMC-Nets7, ERA-Nets8 and Technology Platforms9 provide a good

basis for policy experimentation at the European level, and they should be therefore

continued (European Commission 2008: 11).

7 Acronym for Open Method of Coordination. For more information on OMC-Nets see: http://ec.europa.eu/invest-in-research/coordination/coordination02_en.htm

8 Acronym for European Research Area. For more information on ERA-Nets see http://ec.europa.eu/research/era/era-net_en.html

9 For more information see http://ec.europa.eu/research/innovation-union/index_en.cfm?pg=etp

22 Reflexivity of innovation systems

3.2 Ten quality criteria for reflexive innovation systems

We argue that these capacities should be present in all three stages. In all stages it will

be necessary that actors are at the same time self-reflexive and able to anticipate the

values and interests of others, and there should be the capacities to deal with uncer-

tainties of the present and the future and to react to them. Obviously, some of these

capacities seem to have a 'natural' relationship to some of the stages, for example in

the stage of system analysis self-reflection capacities and anticipation capacities play a

central role. In the stage of strategy development and implementation, certainly bridg-

ing and integration capacities will help to sustain collective coordination, democratic

decision-making and joint action.

Nevertheless, in this conceptual step, we relate the four capacities in a matrix structure

to the three stages (see Table 2) in order to arrive at a model for further discussion. In

this 3x4 matrix, we identify ten relevant relationships which appear to be distinct. There

are no 'blind' cells with impossible combinations, but we merge the three stages into

one in the case of experimentation capacities as this reflects best the understanding of

experiments as learning at all stages and on all levels.

As a result, we define ten quality criteria for reflexive innovation systems which are

described below including illustrative examples of relevant actors and/or policy instru-

ments as well as suitable indicators for measuring the presence of these reflexivity as-

pects. We are convinced that only the empirical application of this matrix can reveal

further insights regarding the relative relevance of the cells. Empirical discussion will

show which of these cells are needed in order to qualify an innovation system as ‘re-

flexive’. We assume that some of these quality features will appear to be very important

for the system, so they might be identified as obligatory quality features. Others might

appear as second-order features which exist in companionship with other features. The

ten quality criteria for reflexive innovation systems are:

(1) Self-reflection capacities in situation analysis

The central questions are: Do the actors critically reflect their own situation? Are they

aware of their values and orientations? And do they consider other actors' situation,

values and positions? The actors in the focus of this analysis should be the relevant

stakeholders in the innovation system of interest. In most cases, these will be organiza-

tions. Potential indicators to measure this level of self-reflection are content analyses of

position papers.

Reflexivity of innovation systems 23

(2) Self-reflection capacities in goal formulation and specification

Central questions are: Are goals formulated at all? Do actors show the capacity to re-

flect about their goals and to articulate them? Have actors taken up general, overarch-

ing goals and specified them to their situation (e.g., "how we can contribute to sustain-

ability")? To what extent are the goals adaptive to their environment, e.g. changing

framework conditions or other actors' goals? The actors in focus are mainly the same

as in the first cell. However, one can also look at this at system level and ask whether

there is a system-wide process to reflect about goals or whether system-wide goals

have been explicitly formulated. Potential indicators are position papers (as in the first

cell), and organizational mission and vision statements. We might also find goal formu-

lation as the first step of (system-wide) strategy and agenda processes.

(3) Self-reflection capacities in strategy development and implementation

Although there is a logical link to at least cell no. 2, the central questions here are: Is a

strategy explicitly developed? Are the goals reflected/applied in the strategy? How

does the strategy address its environment, i.e. the framework conditions? Is the strat-

egy resulting from a reflection/learning process? To what extent is the strategy imple-

mented? Are the success and suitability of the strategy evaluated on a regular basis?

Is the strategy adapted accordingly, if necessary, in other words, do the actors learn

along the pathway? Again, the units of analysis can be the actors in the system or the

system as a whole. Potential indicators are strategy processes and papers, agenda

processes and their outcomes, or formative policy evaluation.

(4) Bridging and integration capacities in situation analysis

Bridging and integration capacities in situation analysis can be analyzed at actor-level

and at system-level. Central questions in situation analysis are: Which knowledge

sources are being used or integrated? Do they reflect different types of knowledge,

different academic traditions and other knowledge sources? Are different points of view

accepted? These might serve as indicators: diversity of members in review panels, task

forces or expert groups, 'new' members in such groups beyond the 'usual' suspects;

communication style in debates: more 'arguing' than 'bargaining' as an indicator for

high bridging and integration capacities.

(5) Bridging and integration capacities in goal formulation and specification

Here, the central questions are: Who participates in goal formulation and specification?

And how is the diversity of interests and goals of the actors addressed – in particular at

the system-level, where consensus cannot be expected? Do we find conflict recogni-

tion and moderation at this stage? Potential indicators are certain process characteris-

24 Reflexivity of innovation systems

tics of goal formulation: Existence of participatory agenda-setting, critical reflection on

power imbalances, and the degree of inclusiveness.

(6) Bridging and integration capacities in strategy development and implemen-

tation

For the stages of strategy development and implementation, similar process features

are needed as in the case for goal formulation and specification. Quality criteria of this

stage are: Are strategy processes interactive? Is strategy-making and implementation

coordinated across relevant actors? Potential indicators are similar process indicators

as in cell no. 5: Heterogeneity of actors involved, coordination across silos, as well as

bottom-up and top-down elements for strategy development.

(7) Anticipation capacities in situation analysis

Central questions to analyze anticipation capacities are: Are the actors aware of poten-

tial future developments? Which uncertainties do they recognize? How do they deal

with them? This can be a quality at actor-level as well as at system-level. Indicators to

look at are: Use of forward-looking techniques such as foresight, radars and scenarios.

We should also look for evidence of this quality in communications and publications.

(8) Anticipation capacities in goal formulation and specification

A central question is whether "futures literacy" is also present in the stage of goal for-

mulation or specification: Do we find openness towards goals for alternative futures?

Does the goal formulation recognize potential uncertainties? Are existing goals recon-

sidered in light of alternative future developments? This quality is present if goals do

not confirm present routines, but help to overcome potential lock-in. Potential indicators

– in particular at actor level – are: Existence of processes to break organizational lock-

in or to open mindsets beyond daily operational work (e.g. transition management, ho-

rizon scanning).

(9) Anticipation capacities in strategy development and implementation

At the stage of strategy development and implementation, similar questions are central:

Are strategies adaptive and robust in light of the identified uncertainties and potential

(unintended) side-effects? And is the use of anticipation capacities systematic, i.e. do

we find a foresight culture? Potential indicators are the existence of systematic use of

foresight and systematic assessments of the environment (e.g. environmental scanning

(360°)).

Reflexivity of innovation systems 25

(10) Experimentation capacities in all stages

Central focuses of analysis are experimentation capacities and actual examples of ex-

periments: Are experiments used to integrate new actors or new goals or to overcome

a lock-in of the system? Is the implementation of parallel approaches allowed for, fos-

tered and accepted? Potential indicators are new formats for discussion or coordina-

tion, or new instruments as an alternative to established ones. An important feature of

such new formats and instruments is that they support a constant learning process,

e.g. a continuous questioning of values, goals and approaches and a search for (paral-

lel) alternative approaches.

26

R

efle

xiv

ity o

f innovatio

n s

yste

ms

6

Refle

xiv

ity o

f innovatio

n s

yste

ms

Table 2: Quality criteria of reflexive innovation systems

Self-reflection capacities Bridging and integration capacities

Anticipation capacities Experimentation capacities

Situation analysis Critical reflexion about val-ues and orientation.

(1)

Indicators: position papers

Transdisciplinarity, collective intelligence

(4)

Indicators: more arguing than bargaining; diversity of stakeholders, new actors

Futures awareness, recogni-tion of uncertainties

(7)

Indicators: reflected in com-munications and publica-tions; processes dealing with uncertainty (scenarios, ra-dars, foresight)

Learning through failure on all levels and in different contexts

Allowing for parallel ap-proaches

(10)

Indicators:

Experimental practices

Goal formulation and specification

Articulation, adaptation - On both levels – individual and systems level

(2)

Indicators: organizational mission and vision state-ments, system-wide agen-das..

Participatory goal formula-tion, conflict recognition and moderation

(5)

Indicators: Existence of par-ticipatory agenda-setting; critical reflection on power imbalances, degree of inclu-siveness

Openness towards goals for alternative futures

(8)

Indicators: Existence of processes to break organ-izational lock-in or to open mindsets beyond daily op-erational work, e.g. transition management, horizon scan-ning

Strategy devel-opment and im-plementation

Articulation, adaptation and learning

(3)

Indicators: strategy proc-esses and papers, agenda processes and their out-comes; formative policy evaluation

Interactivity and coordination

(6)

Indicators: Heterogeneity of actors involved, coordination across silos, bottom-up and top-down elements for strat-egy development

Awareness of uncertainties, assessment of (non-intended) side-effects

(9)

Indicators: Existence of sys-tematic scanning and as-sessments (environmental scanning (360˚)), foresight cultures.

Sources: Own compilation, integrating insights from Voß et al. (2006), Weber and Rohracher (2012), Daimer et al. (2012), Kuhlmann and Rip (2014).

Reflexivity of innovation systems 27

3.3 A revised heuristic of innovation systems

Introducing governance thinking into the innovation system heuristic opens up a new

analytical focus on innovation systems which we have elaborated in the previous sec-

tion. The proposed concept of reflexive governance is a set of ten quality criteria of

innovation systems. As such it is also compatible with the descriptive heuristic of inno-

vation systems. We argue that reflexivity is a quality of innovation systems that should

be present similar to other framework conditions of the system, e.g. such as certain

cultural characteristics or infrastructure.

Translating this conceptual thinking into a graphical representation was an iterative

process during which we drew on existing innovation system frameworks – particularly

Kuhlmann and Arnold (2001: 2) – and developed these further by integrating compo-

nents of reflexivity thinking. Figure 1 shows an intermediate result of this process. The

main subsystems of the innovation system are depicted at the centre (political system,

education and research, intermediaries, industry, and consumers/producers), whereas

relatively stable key factors and basic conditions influencing actors’ behaviour and in-

teraction between the subsystems are represented as different layers surrounding the

core. In addition to general framework conditions and infrastructure, a new ‘reflexivity

layer’ was introduced.

With the aim of integrating the revision of established innovation system frameworks as

proposed by Warnke et al. (2016), we added a reflexivity layer to this innovation sys-

tem model (see Figure 2). This new way of presenting the innovation system heuristic

takes into account several aspects which are also of relevance in our concept: they

introduce new actors such as philanthropists, innovation intermediaries or societal ac-

tors and they present a more adequate and enriched picture of science-society rela-

tions.

28 Reflexivity of innovation systems

Figure 1: Graphical representation of a Reflexive Innovation System (based on

Kuhlmann and Arnold 2001: 2, own compilation)

Intermediaries

Education andResearch System

Political SystemIndustrialSystem

Demand

Professional education and

training

Higher education and research

Public sector research

Large companies

Mature SMEs

New, technology-based firms

ResearchinstitutesBrokers

Consumers (final demand)Producers (intermediate demand)

Government

R&I policies

Governance

Bridging and integration capacities

IPR and information

Taxation and incentives

Refle

xiv

ity o

f innovatio

n s

yste

ms

29

flexiv

ity o

f inn

ovatio

n s

yste

ms

29

xiv

ity o

f inn

ovatio

n s

yste

ms

29

ity o

f inn

ovatio

n s

yste

ms

29

of in

novatio

n s

yste

ms

29

inn

ovatio

n s

yste

ms

29

novatio

n s

yste

ms

29

vatio

n s

yste

ms

29

tion

syste

ms

29

n s

yste

ms

29

syste

ms

29

ste

ms

29

em

s

29

ms

29

s

29

Figure 2: Improved graphical representation of a Reflexive Innovation System (own compilation based on Warnke et al. 2016)

SocietyConsumers, User Innovators

Social Entrepreneurs, Collaborative innovators, citizens

EducationPublic and private educators

on all levels

MediatorsApplied research, Clubs,

associations, trade unions, cluster managers, NGOs

InfrastructureICT, Internet, databases, Co-

Creation Platforms …

InstitutionsIPR, standards, norms

CultureSocial and relational capitalValues, lifestyles, attitudes

Innovation Frameworks

PoliciesPs influencing innovation

framework conditions (RTI) and demand patterns (energy, environment, mobility, health,

defense, home …)

FinancersBanks, venture capital, philanthropists, crowds

Innovation Input

Innovation Supply and

Demand

ResearchUniversities, RTOs, citizen

scientists ...

Public SectorPS actors generating and

demanding innovationCities, hospitals, administrations …

BusinessFirms of all sizes and sectors generating and demanding

innovation

Reflexivity

30 Discussion and conclusions for STI research and policy

4 Discussion and conclusions for STI research and policy

In this paper we propose to further develop and partly revise the current systems of

innovation approach as a response to the inability of this heuristic to convincingly cope

with the challenges of directionality and normative orientation in contemporary STI pol-

icy. In order to better incorporate the requirements posed by this "normative turn", we

introduce a set of four capacities – self-reflection, bridging and integration, anticipation,

and experimentation – that signify the capability of innovation systems to guide innova-

tion processes towards desired ends.

Our vantage point is the critical assessment of the narrow economic growth perspec-

tive in established systems of innovation approaches. While the systems of innovation

heuristic remains a valuable conceptual frame for the analysis and the design of STI

policy, it falls short to address issues of directionality and questions which innovations

are desirable from a societal perspective. In the recent past, fixing market failures and

systemic imperfections as the dominant rationales for economic and innovation policy

have increasingly been complemented by new rationales for state intervention and in-

novation policy (Foray et al. 2012; Mazzucato 2013). Enacting these new rationales or

'missions' requires additional and different approaches than those solely aimed at fos-

tering generic innovation system functions such as interactivity, mutual learning or dif-

fusion. In effect, supporting certain technological trajectories and innovation paths ex-

plicitly entails the partial departure from the conventional principle of 'neutrality' (Aghion

et al. 2009: 688).

Due to its focus on strategic interaction and on actors' intentionality to influence the

quality and direction of innovation processes, the integration of a dedicated governance

perspective into the systems of innovation approach is a promising conceptual vehicle

to support the identification and assessment of substantive goals, and eventually the

guidance of decisions and actions to achieve these ambitions within a given innovation

system. Of the rich theoretical and conceptual literature on the governance of STI, we

conclude that the literature on reflexive governance is particularly useful in this context.

While normative directions for the governance of research and innovation, as widely

articulated in missions such as the so-called ‘grand challenges’ and in part promoted

by the RRI discourse, are plentiful, these normative directions are – by and large – im-

posed upon the innovation system actors exogenously, and neither their broad accep-

tance nor their adequacy are a given. With the aim of increasing the sense of 'owner-

ship', legitimacy and eventually societal robustness, we propose to partly endogenise

processes of defining and specifying normative directions. Putting increased emphasis

on and improving the capacities of reflexive governance will potentially contribute to the

Discussion and conclusions for STI research and policy 31

strengthening of the innovation systems' input-legitimacy. As such, the often criticised

preoccupation with output-legitimacy (i.e., the economic performance of innovation

systems) can be rebalanced in favour of a stronger focus on the input dimension.10 Of

course, the formal mechanisms of legitimacy in the context of representative democracy

remain in place. But in view of the specific characteristics of the STI field such as a high

degree of autonomy and self-reliance of the actor landscape together with the substan-

tive openness of research and innovation directions, hierarchical steering and conven-

tional governance settings seem even less suitable in this field than in most other policy

areas. Within this setting and in view of the requirements that need to be fulfilled to facili-

tate innovation systems' self-reflective capacities, to orchestrate the opening-up and clos-

ing-down of decision-making processes, and to implement appropriate policy and regula-

tory measures, the state will have to play a particularly decisive role.

Limitations and future research

We are convinced that the proposed reconceptualisation of the systems of innovation

heuristic along the lines of intensified reflexive governance, operationalised by a set of

four capacities and related quality criteria, will potentially improve the analysis of inno-

vation systems' ability to address the complex challenges of directionality and, in the

long run, also the